Improving Extreme Learning Machine Performance using Ant Colony Optimization Feature Selection. Application to automated medical diagnosis

Abstract

Extreme Learning Machine (ELM) is a single-hidden layer feedforward neural network, where the weights between the input and hidden layer are randomly generated and never updated, whereas the hidden-output weights are analytically computed. Theoretical studies have shown that ELM maintains the universal approximation capability. Artificial Intelligence applied in automated medical diagnosis is problematic due to the high risk of overfitting the data, because of the large number of attributes. The goal of this paper is to propose a feature selection (FS) mechanism based on Ant Colony Optimization (ACO), in order to speed up the computational process of the ELM. The proposed model has been tested on three publicly available high-dimensional datasets.